The vast majority of tuberculosis (TB) cases occur in resource-limited settings and despite potentially curative therapy, deaths from TB outweigh any other single bacterial pathogen. The highest rates of mortality and greatest consumption of resources reside with severe forms of TB disease, including multidrug-resistant (MDR)-TB and the neglected disease states of pediatric TB, TB meningitis and TB sepsis. Pharmacokinetic variability (suboptimal circulating drug concentrations) appears increasingly common and is particularly actionable through dose increase but has not been studied in multisite prospective cohorts or for the most severe forms of TB disease. Additionally, conventional drug-susceptibility testing for TB is crude: it uses one concentration of drug and therefore does not allow for quantification of resistance. Our preliminary work suggests quantitative susceptibility will better inform drug choice and can be adapted to fieldable molecular diagnostic platforms to ultimately deliver high-yield, individualized and actionable results for the most burdensome of TB disease. This proposal will unify established collaborators from 4 sites of diverse TB context (Tanzania, Uganda, Bangladesh, Siberia) to strengthen laboratory resources, develop personnel and build research infrastructure to study prospectively the extent and mechanisms of pharmacokinetic variability and M. tuberculosis drug-resistance (MIC) and their dual impact on treatment outcome. Over half of this ICIDR funding goes to the international collaborative sites. This proposal can contribute immediately to regional algorithms of TB treatment and will also build broad capacity for infectious disease molecular diagnostics and field epidemiology.